Transcripts
1. Introduction:
2. Midjourney Setup:
3. Midjourney Plans:
4. Midjourney Model Versions:
5. Midjourney Settings, Presets and Options:
6. Prompt Guide: In this video, we are going to discuss about the things to include in prompts on how to make sure that we
get that result. When typing prompts try to be
as clear as possible about any context or details that you feel are important to you. Think about subject,
media, environment, lighting, color, mode,
and composition. These details not only helps me journey to
produce the art, but also helps in generating
the desired image that exactly what you
expected it to generate. Please consider
pausing this video and going through the retails
mentioned in the video. Or take a screenshot so
you can reference and keep in mind when
generating your artwork.
7. Prompt Styles: In this video, we are
going to discuss about prom stands with magenta. You can easily create an
art through a single bond, but it will be based on a
journey is default set. However, you can easily
modify that based on design style that you like
or want in our artwork. You can consider
design cells such as artistic sketch
style or historical. Styles, include
different locations and environment in
your art design. Let's see few examples so you get better
understanding of it. Here are few examples
of artistic styles. There are two examples
of sketch styles. Here are a few example
for history styles. Here's few examples with
different environments. You can even add different emotions in your
designs such as these. We can also add
different color theme in an image in which we could describe
particular or a set of colors, such as this. Let's try to create a new image from each exam by different examples that
I showed you earlier. See what Jenny proteases. We are going to include border, color from artistic style team, 1950s from historic
style, theme, pastel from color, theme, and better from expression. And at last, we are
going to include cloud forest as our
background or environment. As you can see that
magenta has produced and unique style from
those combinations. In future video, I'm
going to discuss in-depth about a topic called
multi prompts, where you will learn
how you can give importance to certain
elements in an image. So Magendie produces exactly
what you want it to. Gender.
8. Prompt Parameter Guide: Hi guys. In this video we're going
to discuss in-depth about my journey prompt
parameters are first, we are going to discuss
batch desk ARR aspect, which is the aspect ratio of your ML that magenta
it generates. By default, the aspect
ratio of my journeys are nice to have it is
in square shape. However, you can change
that based on your needs. E.g. we can choose to use
three is to two aspect ratio, which displays in horizontal
or landscape images. Or we can toast you, do is to three f patricia, which displacing vertical
or portrait images. Let's look at how we can
use this parameter now run. As usual, we're going to
type in forward slash. Imagine. Then our prompt. Besides that, we are
going to add dash, dash, ER or aspect. These to do. Since I wonder
imagine horizontal. Let's look good that is. And see how it generated. As you can see, the
generated images are in three-to-two
aspect ratio. Hope you get the idea and implementing our
image generation. Up next, It's dash, dash, test and test B. These are actually made
journeys testing models. I want to show you
guys the difference so you can better
understand them. So we're going to type in
both the parameters beside our prompt and see what
the results may generate. And as you can save the test, it sort of looks more
like an artistic style. Whereas with that SP, the images, it looks
more photo-realistic. So if you want so sort
of like a stock images, you can try it SP
and see results. Another thing to notice
with test and SP on the donee supports today's to-do and do is to three aspect ratio, which is the max that
you could choose. And you will get
only one image as output when you use it
without any aspect. But I've entered, that
is if you use default, the aspect which
is one is to one. But if you use Aspect Ratio, aspect parameter, you will get a result of two outputs
or image generation. Up next, we're going to discuss about dash,
dash creative, which can only be used to
besides test and SP parameter. We will try by using it with and without test
parameter and see the results so that we can get a better
understanding of it. So as you can see,
this is a result generated with the parameter. As you can see, the image
has an artistic style to it. Whereas the creative
parameter, besides the test, we can see that magenta has taken control over the
creativity and added, but that looking
details to the images. So that's the difference
between NADH using creative and using creative
parameter bizarre test and SP better emitted up
next is dashed as stylized. We could use this to
stylize our image, or you could say to make our
image more extraordinary. The default way in a way from different
Magendie versions. However, since we are using
a journey for the range for it is 0-1 thousand and
the default value is 100. The more the number of value is, the more unique and different
it will be from your, from like these examples. This is how you can use
this parameter besides your prompt by typing
in dash, dash, yes or stylize and value and
extra training 0-1 thousand. Alternatively, you can even set the default style option for your image through
forward slash settings. Don't have to keep typing in the parameter everytime
agenda to image. Up. Next, we're going to
discuss about dash, dash. Now, with this parameter, we can eliminate
negative elements that we don't vote in our image. This parameter as usual, by typing in forward slash
imagine, and then prompt. Beside it, it will
be dash, dash. And next to it will
be the subject that you don't want
in your images. In my case, I don't want to include the tracing the image. So let's see what my
journey produces. As you can see from
the results that are almost non-existent
except few bushes, which we could
easily eliminate by using photo editing
software such as Photoshop or free alternatives
such as gamepad, Krita. Up next we're going to
discuss about dash, dash q, which is about
the quality of an image. By default, the quality is
set to one in my journey for. However, on the older versions
you can go up to two. Note that the higher the value, the more time it takes to generate any maze by my journey. We can use this in our problem. We're going to
type the prompt as UDL can be that it will be dash, dash q and the value next
to it will be running 0-1. Here's what, Jenny,
for different quality. The left image that you
see has a quality of one and the right image that you see has a quality of 0.25. As you can see, the
intricate details and high definition is more in quality when compared to quality zero point to fight. So I hope you guys get the
idea on how to use this. Dislikes to analyze. You can also read
the default quality for you and image generation through for restless setting. Now let's get to know about
does that say parameter. They say they're
especially useful when generating an image with
similar style or theme. E.g. we want this image in human art form rather
than as a reward, but with similar or same. Stan. Let's see how it works. First, we are going
to add a reaction to the image that will
send by my journey. Where we can see many images. We'll scroll down to objects and find an envelope emoji that is covered or simply we can just type in such as envelope. You can easily find it and
then you need to click on it. My journey then
sends the message in private message the details
regarding the image, which also does eat
information that we need. We are going to copy that and come back to send
message to meet journey. We're going to type
in our ramped using forward slash imagine and
then our text prompt. And we said it will be dass die seed and extrude that
seed number value. By entering magenta
will produce a result. And let's see what
it comes up with. As you can see on the left, we got the image
of a robot phase, whereas the image
on the right side has a human face to it. The style and the
theme is almost matching to that
of an afterimage. Hope you get the idea
and you can utilize it. Feel free to work around and try with some of your artworks. And I'm sure you will feel mesmerized by using
this parameter. The next we are going to
discuss is dashed as chaos RC. With this parameter,
you cannot omit journey to produce an image
with chaotic behavior. The chaos value ranges 0-100 and the default
value is zero. We are going to use it beside
our prompt using dash, dash, dash, dash chaos
followed by zero to 100 value. The more value we choose, the more chaotic behavior
will be in an image. Which means that
the image will have more unexpected or
unusual results. As we can see the results, the more value increases, the results are different. The chaos of 100 as the results
that look for different from what kills off zero
or 50 does look similarly. You can see in this
result that all looks unusual as we go on. I have chaos. Where do you these were all basic parameters
that you need to know to get started
without our generation. Up next, I'm going
to discuss a lower, I'd once parameters that Gordon non-zero image dendrites
and capabilities. So let's go and do that. And I hope you guys
understood all the basic than a meter that
I experienced so far. If you still have any doubts, feel very Dawson Qian ethics
and send me a message. I will never looking at
those and help you out.
9. Advanced Prompt Guides: Hey guys, In this lecture video, we are going to discuss
a few advanced problems that could help you improve
our demand generation. In this video, we are
going to discuss how we can create cool
images through emails, prompts, and how you can improve
and work around with it. So to use an image on Discord, we're going to click on the plus button in the
message bar on discard. And then we are
going to click on Upload File and choose
our desired image. And then after that, you can
open that image and click, Enter on the uploaded image, then copy the link
of that image. After copying, we're going
to type in forward slash. Imagine there's that
image link and write some text prompts so we could customize it according
to our ideas. I'll try this with neon, my brand color effect and see what Magendie produces
results for us. As you can see,
we got the result that made DNA produced based on our image and text prompt that map scale the first one
and see it more clearly. Here is obscured, embed, it seemed squared cool. You can feel free to try
it with different and unique text prompts and see what results you can
generate with magenta. Up next we're going
to try blend mode. With magenta. You can produce a
combined image, 2-5 images like
you see on screen, the combination of two images is Missouri are combined into one and the new piece of artwork is produced
by mid journey. Similarly, you can
see on the Aardvark beds three images that
are combined into one. This process of combining can
also be called as blending, hence that I'm blend. Let's see how we can do this. We are going to use
these two images. We're going to type in
forward slash and then blend by typing
by, I mean, sorry, by clicking Enter, you will
get the options to upload minimum to images that may select those two files
and upload them quickly. There are other options as well. Let me go through quickly
and before proceeding. As you can see, we can specify the dimension of output and that you are going
to receive from a journey. And we've got portrait
mode, square, landscape. So you can choose any one
of these out of three. Perform this. I already
mentioned another one which is you can upload up to five images and
combine them into one. And let's see what Magendie producers from those two images
that we uploaded. This is what Magendie produced. You can try this
blend feature and see what you can achieve with the combination of your artwork. Up next, we're going to discuss about the mixed mode art prompt. You can turn on this
remix more Bye. Typing forward slash settings. And by enabling it, you can get the option
to make variation with similar or
same image teams. With trimming from,
you can easily change prompts and parametered
soften image. This will take the
gender composition or reference of air initial
image and use it as a reference to create a
new variation of it by adding in more
details in the prompt and changing its aspect ratio. Here's what we got with
the help of rhythmic mode. Now we're going to discuss the most useful one
that is multi-pronged. As you can see on the
left-hand vitamins, both can be called as hot dogs, but they both have different
depth and meaning to them. You, when you are
generating an image, they add eats and
understand the prompt as one line and
separates them based on the prompt
characteristics that you add through commerce while creating. You might think of
something but edge and it's different results
in such cases, you can be more specific about that by adding double
colon in a prompt. This helps understand that those are two different words
with different meanings. Like in this example
on the left image, the cupcake is considered as a single word in
an illustration. Hence, it's displaying cupcakes. While on the right image, you can see that the
cake and copper, two different subject
in the illustration. Since we have used double
colon in the middle, That's where my journey is
displaying cake inside a cube. Let's see how we can use
this multi-pronged feature. In our illustration. At first, we will normally write it as
a single word and submitted. And on the second time
we will submit it with double colon and see
what Magendie produces. As you can see, we got our
results from a journey. The first prong we typed in
without the double colon, just the dog word. As you can see, there's
just a photo of our dog. And here you can see the second
one that looks more like hard as in effects
in a dark image. Hope you guys get the idea and understand the concept
of multi-pronged, by which you can do exactly what you wanted
from my journey. Up next is prompt waits. With this prompt, you
can give importance to certain elements of your
MS based on your choice. You can use this immediately next to double colon,
something like this. You can give it any
value to this diagonal, understand the precedence of it. However, if you do not
specify the weights, it will have always one
as its default value. And it doesn't matter
what font weight value. But you must make sure
that the prom value, that EQ on the one. So this should be greater
than the other, like e.g. as you can see on the display, the value on the
one side is one, whereas on the other side, It's. Now, as I mentioned earlier, that you don't give any value, it will be set to
default as one. So this will result
in one on each side and will not produce any
change in your results. Likewise, you can find the similar scenarios
we spread on the image. You can also add a
decimal points like decimal numbers like
0.0 by 1.5, etc. You can also add
negative value to them, is you don't want something
to run into our image. It works similar to the dash, dash know that I told you
in previous lecture review. But instead of completely suppressing the
subject or element, it will try to suppress it or do less importance based
on the value that you need to make sure
that the total value of all the weight
values should be equal to the positive number or not lead to any
negative number. Let's try with the prompt
that I despaired and see what results we
get from a journey. All you can see that
regard our results here. I'll add in black as a gift the problems weight
of four for the black color, which means the precedents
are more importance is given on the black and white, most of it Florida
flag that 19 cheered. Do it. So it's like a mixture of gray and white since we gave a negative value to
it. How are you guys? Guard all the prompts proper me I mean, I'm
disrupt properly. If you still have any doubts
or any clarifications, please feel free to us.
10. Upscale the Image Further: Hey guys, Now that we
got to know about me, Jenny and everything and Illustrator will
continue further with scanning the image and adding more details and
make it more refined. As you can see in the image, the image is not so accurate
and look so looks blurry. As you zoom in you when the
image size is competitive. So you can increase the
size by upskilling IT further through solutions that
provide further upscaling. They are both paid and
pre-survey services available, so we'll try out both and see which one
suits for your needs. So as you can see, this one is the first
one that we that is most entrustable and also provides a backup solution. I mean, yeah. You can see the
details are more, much more defined and
much more accurate. As you can see, even more grasses are added or urine compared to the original
image that we uploaded. We can even further up in Kelly four times and see
what results we get. As you can see in this, it's much better. This time. Some details over
here messed up, but you can improve
that if you have Photoshop skills to rise. One of the best solution that you can use
the scale for that. Let me talk about
the pricing of this. This is three. This can be used
for free as well. You can get up to 40 factor. It is permanent and if you
want to use it for that, you can apply it to their
subscription model. As you can see here.
Or you can upload up to 25 and to define pixels. Further than that,
you can upload the image if it's
about that resolution. Now, got the next
one is image to go.com with this also
again upscale every image. This is showing
you the interface. You can click on
that scale image and choose the image
that you want to upload. Like e.g. I want to
click on it, upskill. And as soon as it's uploaded, you can click on start. It will start
upscaling that image. These things take some time. So you need to
have some patients while it's doing its job. Let us go through
it's bland event. This is also a free software. I'm a free website, but there's a
certain limitations. You can go through this. You can apply it to their plan and get
more better results. Since to an audit done? Yeah. Notwithstanding, I guess
let's download it and see how it's looking compared to the one
that we already have. This is original one, and this is the one
that we got high. As you can see in this, I have zoomed in my spot here. And you can see that we can
zoom up so much clothes even with fulsome much in
this general one. So that's the difference
between these two. Yeah. You can try this also. If you are on a
budget and don't have money to spend on upscaling
additional x squared. If you have somebody you
can invest in websites. I'm in software such
as the epithelia. This is the best one by far
when it comes to upskilling. And you can also read reviews
and everything online. It has best reviews. It's yeah. You can see here
the images are much more refined when compared
to the original one. As you can see in this example. If you wanted to buy this,
it is a bears John lifetime. They will only charge you once. I've done that donors
your offer that you can use it for lifetime. You can also try a pre-trial and see whether it is based
on your requirements. I have already installed a
free trial and let's see, a demo set out late work. And yeah, one more
thing to both, the free, free one. You can only upload
one image at a time. Whereas through
gigapixel, you can upload more than
one simultaneously. Android, as you can see, this is an interface
for V0 pixel. You can upload the image
that you want to obscure. Skin, sorry. You can
tweak the settings. Whereas with the stool
that I mentioned, you can't do sustained
benefit too. And you can upscale a CPU or a text or even more, I guess, with this C11 and
benefit this request. Additional files to model. So let it download those files. Or image has been can I guess. Let me see. Yeah. As you can see, as we zoom
in more on the original one, it gets more pixelated. The scale one through
gigapixel, it is, it is having more refined
detests difference. We can see the original image, the rock as pixel. One. Pixel. It gets quantified and
smooth everywhere. So this is definitely
helpful and useful for you and
upscaling your images. You might not notice this
difference when you zoom out, but it does help
you when you like. I use it for printing purposes or for any other
commercial purposes, I will attach a link
for gigapixels, also, the other two EPS of
three objectives that I mentioned in the
description for you guys. You can directly click on
the link and Chiquita.
Dheeraj
